Design and Prototyping of a Tendon-Driven Continuum Robot for Space Debris Capture

Active debris removal requires capture devices able to engage objects whose shape, size and condition are only partly known. A sample of fifty catalogued intact objects drawn from the ESA DISCOS database spans more than two orders of magnitude in mass and five families of primary shape, and three satellites out of four carry deployed appendages: the geometry of the target, more than its mass, determines how a capture mechanism must engage it. This thesis investigates a tendon-driven continuum robot as a compliant capture device that adapts its shape to the target, from modelling to a full-scale laboratory prototype. The manipulator is a single section 2,1m long, actuated by three tendons routed through 20 three-armed vertebrae, with a jointed backbone that retracts into the base and extends along its axis. A constant-curvature kinematic model of the extensible section, including the discrete routing of the tendons and regular in the straight configuration, is implemented in a verified Python tool with forward and inverse simulators. The statics of the loaded backbone is studied by adapting to the single-section arm a benchmark implementation of five modelling approaches, which brought to light four defects in the reference code, and by describing the jointed backbone through an equivalent bending stiffness; the analysis also shows that the slender arm cannot support its own weight under terrestrial gravity, a constraint on ground testing that does not apply in orbit. A multibody model in MSC Adams, built on an earlier configuration of the arm, defines the dynamic analyses and the simulation of the orbital environment. The prototype comprises 3D-printed vertebrae, coaxial intervertebral springs, four stepper motors and a two-level control system operated from a computer or a joystick. Its first assembly revealed a severe instability; a single-interval test traced its local component to the lateral deflection of the springs under the compression of the tendons, consistent with the buckling criterion for helical springs and a failure mode that axial spring models cannot represent. The analysis of this mechanism, the strategies for stabilising the arm and the experimental campaign needed to identify the model parameters are presented as the basis for the next design iteration.